← back
Vibes won't cut it — Chris Kelly, Augment Code
Takeaway
AI accelerates code generation but production software engineering — changing systems safely at scale — still requires human judgment, context, and rigorous practices.
Summary
- Pushes back on AI hype claims that engineers will be replaced; argues LLMs generate code well but code isn't the job — making thousands of design decisions safely is.
- Vibe coding (accepting AI output without review) cannot handle production systems with four-nines uptime, gigabytes of data, and snowflake idiosyncrasies.
- Recommends 'the best code is no code'; every generated line is debt; pattern-matching breaks down at scale where bespoke architecture choices dominate.
- To make codebases AI-friendly: documented standards, reproducible dev environments, easy testing, type systems, and rich codebase context (Augment's pitch).
- Notes professional engineers have been unusually slow to adopt AI compared with prior shifts like cloud and DevOps.
ai-codingsoftware-engineeringproduction
Original description
What's the role of vibe coding in a production-grade applications? Join Augment Code's Chris Kelly as he talks about the role of context in software engineering, not code. About Chris Kelly Chris is the head of developer experience where he works across Augment to make building software better for every developer. He’s been making developers happier and more productive for 15 years at innovative companies like New Relic, GitHub, Salesforce, and FireHydrant. You can find him at @amateurhuman everywhere on the internet. Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter